how to create an alert if the server is shut down

[Arunkumarbnp]

how to create an alert if the server is shut down

I've created an alert for the metric builtin:host.availability.state but received the server down alerts for 2 servers instead of 11 servers.

[AntonPineiro]

Hi,

By default Dynatrace already alert you when shutdown is not grateful.

Check infrastructure anomaly detection settings in environment level, host group and host.

Best regards

[Arunkumarbnp]

Hi Anton 

Thanks for your reply.

I need the alert to be created for the servers that are shut down on both planned and unplanned maintenance 

[AntonPineiro]

Hi,

If they are shutdown gratefully, I would suggest you enable that option because by default, it would only alert you when it is not grateful shutdown.

You can configure this in tenant leve, host group or host depending of your scope.

You do not need to create a metric event about this.

Best regards

[Arunkumarbnp]

Hi Anton

Thanks for your input.

I didn't get your suggestion.

Could you please elaborate on how to set it at host group .

Sorry am new to dynatrace.

[AntonPineiro · Accepted solution]

Hi,

You have more than one way, but you can do it from host view, click on "..." and select "Host group settings":

Select anomaly detection settings in infrastructure menu, and choose "Alert on grateful host shutdown":

And save changes. This configuration would be applied to all host under that host group.

Best regards

[Arunkumarbnp]

Thanks for your reply.
after changing here, do we need to make changes in alerting profile for email notification

[AntonPineiro]

Hi,

You are welcome.

It depends how you have created your alerting profile. But yes, you need to be sure problerms are matching your alerting profile.

Best regards

[Arunkumarbnp]

sorry anton, 
can you please mention how this alert can be captured in the alerting profile.

I've added the event filter in the alerting profile as Predefined: Contains events of type 'Host gracefully shutdown'

would that be the one to trigger alert?

[AntonPineiro]

Hi,

It would depend about what to would like to filter in or filter out in your alerting profile. Do you want isolate this event?

A good practice would be define your host groups, define your management zones, and you can create alerting profile linked to a management zone.

But if you want to isolate only that event, I would suggest you force that problem, and checking problem title and Davis categorization (impact level, severity).

And base on that, you can create your rules and seeing if problem is matching your alerting profile or not.

Best regards

[ChadTurner]

As @AntonPineiro stated. There is a hierarchy for the "Alert on Graceful shutdowns" and long with alerting on non-graceful shutdowns ... Meaning someone pulled the plug. 

The hierarchy is as follows starting with the most granular level: 

1 - Host by Host within the host settings. (Only applies to the host you set it on) If not set, then see #2

2 - Hosts by Host Groupings within the Host Group Settings. (Applies to all the Hosts in your Host Group) If not set, see #3

3 - Environment Level. (Applies to all the Hosts in your Environment).

Also, Keep in mind, if you have a host in a Maintenance window, planned or not, and it is set to ignore alerts and not notify, then no matter what you have set, it will not alert. My recommendation is to not only verify the alert settings by way of what @AntonPineiro and I have outlined, but also ensure the hosts that didn't alert, were not part of a maintenance window at the time of the shutdown event.